Skip to main content

Full text of "Duty trial of a pumping engine for the Louisville Water Company, Louisville, Ky."

See other formats





Louisville Water Company, 






The William Cramp & Sons 

Ship and Engine Building Co., 



of Experts on Engine No. 3 

To the President \md Directors Louisville Water Company, 

Louisville, Kv. : 

Ge n(le men — In accordance with your request, the contract trial 
of Pumping Engine No. 3 was made under our direction during the 
six days beginning at 3 p. m. April 25, 1894, and ending at 3 p. m. on 

May 1, 1894. 

Description of Plant. 

Pumping Engine No. 3 was designed by E. D. Leavitt. of Cam- 
bridgeport, Mass., and Charles Hermany, of Louisville, Ky., and 
built by the I. P. Morris Company, of Philadelphia, Pa, It is of the 
Leavitt type, having two vertical inverted cylinders, the piston rod of 
the high pressure cylinder being connected by a link to one end of a 
beam, and the low pressure similarly to the other end of the beam. 
The main shaft is at one end of the engine, and the connecting rod 
passes from a pin in the upper part of the beam to the crank pin. 

The steam pistons have opposite motions in consequence of this 
arrangement, and the exhausts from the ends of the high pressure 
cylinder take place to the corresponding ends of the low pressure 
cylinder. The steam cylinders are jacketed all over with steam of 
boiler pressure, and the reheating receivers, of which there are two, 
use boiler pressure steam for the reheating medium. 

Each steam cylinder is provided with four grid-iron valves oper- 
ated by Leavitt cams. The point of cut-off in the high pressure cyl- 
inder is automatically determined by a centrifugal governor, while 
that on the low pressure cylinder is fixed. 

The construction of this engine is of the most massive character, 
the weight being far greater than any other pumping engine of the 

same capacity. 

The pumps are located directly under the engine, and are con- 
nected to the beam at such points that while the stroke of each steam 
piston is 10 feet, that of each pump plunger is 7 feet. The plungers 
work vertically and are of the differential type, so that each plunger 
draws water through the suction valves on the upward stroke, dis- 
charges it through the discharge valves on the downward stroke, and 
expels one-half of it into the main on each stroke. Each pump i- 
therefore single acting on the suction side and double acting on the 

4 Report of Experts on Engine No. 3. 

The pump valves are small and have 5-16 inch lift. There are 143 
suction valves and 124 discharge valves in each pump. 

The engine is provided with a surface condenser. The air pump 
is vertical double acting, and is worked by a crank on the end of the 
beam shaft. 

On account of the rise and fall of the Ohio River the bed plate of 
the engine is placed above the highest probable high water mark, 
while the bottoms of the pumps are sufficiently low to draft water at 
the lowest stages of the river. The distance from the bottoms of the 
pumps to the bottom of the bed plate is 61 feet. 

The engine is provided with three boilers of the Belpaire locomo- 
tive type, built by the I. P. Morris Company, having a diameter of 
shell 82 inches inside of the smallest ring. The boilers are fed by a 
Worthmgton duplex pump, the exhaust of which passes into an open 
heater. The feed water is that from the condensed steam of the main 
engine, there being provision for making up with cold water any 
waste that may occur. 

By the terms of the contract between the Louisville Water Com- 
pany and the I. P. Morris Company, it was specified that 

-The trial shall be commenced by taking the engine in its current 
working ^condition, while running and pumping the ordinary daily 
supply of water into the reservoir; the boilers, however, shall be 
elean, and all appurtenances of the machine in proper working order 
U hile in this state of service the engine shall be stopped at about 10 
ocock in the morning, the fires hauled, . the ashpit, furnaces, and 
boiler-room cleared of all fuel and combustible, and all steam blown 

levelThl fi n T' 1" Water therem br ° U * ht t0 the ~I 
level The fires shall then be started fresh, and all the fuel used 

thenceforward shall be weighed and charged to the engine; the wood 

used to start the fires being rated at one-half its weight as coal 

'As soon as sufficient steam has been generated to turn the engine 

over agamst its regular water load, it shall be started and run, un Z 
rruptedly, for one hundred and forty-four hours, and stopped „ 

the seam pressure m the boilers shall have been worked Ln 
paint a winch the engine will no longer revolve against its load 411 

he coal and other fuel consumed during these one hundred and forf v 
four hours shall be charged to the engine in its merchantable f on 
and condition, and no picking out of inferior coal, or ded ,on 
ash, clinker, cinder, or incombustible of any kind ^tf^Sj 

B !ii i in. L ii. i \Y 


* ■ 

I trial n 

a ' uri ! 

Pit! ira and dm ren 

houi l imberland bi mil I 

the puni] 1 1 r ill 

h lit I ii i i tl pum| 

in t, vvl 

irn i mntei 

I >( J I M | 1 |M| I f>|| 

K\ nd hi 


< p l hit . I i 

1 1 1 I \u ' m ad i p t n d i \ 

in this Hid during p 

into hi« Ii i If i I I, 

deli in then r 1 1 nil 

I ns ot Wei hunt* ol il 




1 t 

o ye i i it the 

ii ; i d t ) I • I i I 

\i 1 1 | H'l I III 1 1 I | m| I 

Itll l I 

ik I I foi In i |M 

i i 



>l * 

I I 


I > I It) 

I . I 
in ' ir 


; I I l ' 

1 1 \i 
id ( ii |" 

!. The I I .<. 

A l . ■ • 




ir I 


rhe al i 

Il v m I 

,u il 




• < 



I V 



6 Report of Experts on Engine No. 3. 

new fires started. As soon as steam began to rise the engine was 
started. Steam pressure was increased to the working pressure of 
140 pounds, and the engine speeded up to about 18.6 revolutions per 
minute as rapidly as possible, and run 144.17 hours continuously. 

Boilers Nos. 2 and 3 were used. 

The coal was weighed in barrows as needed, separate accounts 
for each boiler being kept. 

Samples of the coal were taken daily, and the amount of moist- 
ure ascertained by drying for 24 hours on top of the boiler flue. 

The main feed water was weighed on accurate scales, and the 
jacket and reheater condensations were returned to the boiler by 
gravity through a Worthington meter. This meter, although dealing 

with water of 328 F. temperature, worked with perfect uniformity 
throughout the run. At the end of the trial it was calibrated under 
the working conditions by a run of three hours. 

The feed water was weighed before passing through the feed-water 
heater, and the amount of water added to the feed by the condensa- 
tion of the feed pump exhaust was computed from the rise in tempera- 
ture of the feed caused thereby, and from the work done bj the 
pump. The heat radiated by the pump and its steam pipe was neces- 
sarily ignored. 

Indicator diagrams were taken every hour throughout the trial, 
and readings of the engine counter, steam pressures at boiler and 
engine, vacuum, barometer, force main guage, height of water in well, 
temperatures of air, steam, feed water, and escaping gases from flue 
were taken every thirty minutes. 

All of these readings, as well as the weights of the coal, were 
taken simultaneously by two observers, representing both parties to 
the contract. 

The scales used for weighing the coal and feed water were adjusted 
and tested on the day preceding the trial. 

All thermometers and gauges were carefully tested. The head on 
the pumps was determined by a mercurial column connected with the 
pump chamber just above the delivery valves. To the reading of 
this gauge was added the distance from the zero of the mercury col- 
umn to the surface of the water in the pump well. While the friction 
head through a portion of the pump and the whole of the delivery 
mam is thus credited to the engine, that of the water through the 
lower part of the pump and pump valves is not. 

A calorimeter was used on the steam pipe near the engine and a 
separator removed the water of condensation from the steam pipe. 

8 Report of Experts on Engine No. 3. 




5. Diameter of L. P. piston rod, 6 ]rL 

6. Stroke of each piston. IO f t 

7. Mean clearance H. P. cylinder, 1-585 per cent. 

8. Mean clearance L. P. cylinder, i-53° P er cent. 

9. Diameter of each differential plunger, 34 in. and 24 1 16 in. 

Stroke of each differential plunger, 7 ft. 

1 1 . Ratio of steam piston areas, 4 to x 

Volume displaced by one stroke of one plunger, 330.15 gal. 

13. Volume displaced by plungers during one revolution of 

. en S ine > 660.30 gal. 

14. Diameter of each discharge pipe, 24 j n 


Dimensions of the Boilers. 

J 5- Type— Belpaire Locomotive, Double Furnace. 
16. Number in use, 


2 in. 

17. Smallest inside diameter of .shell, g 

18. Inside length of fire box, 8 ft. 1% in 

19. Inside width of fire box, twice 3 ft. 9 in , 7 ft. j n 

Length of combustion chamber, 7 ft 

Length of tubes, 16 ft' 

22. Diameter of tubes outside,... 

23. Number of tubes, each boiler, J 

7 ft 1 ,„ 

25. Width of grate during trial bricked up in each boiler to 5 ft * in' 

26. Heating surface, each boiler, a , 4rv J f ' 

27. Grate surface during trial, each boiler, ^ l8 - sq ft 

28. Ratio of heating to grate surface during trial, ^60.24 to 1 

Length of grate, each boiler, '''''///////. 1 

I ; \ >. i i u i \ i 

io Report of Experts on Engine No. 3. 

59. Total water pumped into boilers,. 991,518 lbs. 

60. Total water returned to boilers from jackets and reheat- 

ers, 189,795 lbs 

6r. Total steam used by calorimeter, 727 lbs. 

62. Total water drained from separator 23,428 lbs. 

63. Total moist steam used by engine and feed pump... 1,157,158 lbs. 

64. Percentage of moisture in steam after leaving sepa- 

rator j 0.55 per cent. 

65. Total dry steam used by engine and feed pump,.... 1,150, 792 lbs. 

Steam Used by Engine. 

66. Total moist steam used by engine only, 1,133.768 lbs. 

67. Total dry steam used by engine only, ' ^27,533 ^ DS - 

68. Total moist steam passing through cylinders, 943>973 lbs. 

69. Total moist steam passing through jackets and re- 

heaters, '89.795 lbs. 

70. Percentage of moist steam used by jackets and re- 

heaters, 16.74 pe r cen t . 

71. Moist steam used per hour per I. H. P., 12.223 lbs. 

72. Dry steam used per hour per I. H. P l2 . 156 lbs. 

73. Dry steam passing through cylinders per hour per I. 

XX * ± •> to. 1 20 lbs. 

74. Moist steam used per hour per pump H. P., '3-123 lbs. 

75. Dry steam used per hour per pump H. P., 13-050 lbs. 

76. Prevailing point of cut-off, H. P. cylinder, 20.20 per cent. 

77. Prevailing point of cut-off, L. P. cylinder, 42 10 per cent. 

78. Ratio of expansion, by volumes, 20 4Q 

79. Steam accounted for by indicator at H. P. cut-off, 7.75 

lbs ->— 76. 58 per cent.* 

80. Steam accounted for by indicator at H P. release, 0.166 

lbs — 
u&,,— -^o-S? per cent.* 

81. Steam accounted for by indicator at L. P. cut-off, 10.008 

■ > ' 99.60 per cent.* 

82. Steam accounted for by indicator at L. P. release, 9.7 

11 ■* ' 

'~ 9 6 -o9 per cent.* 

2 5 

British Thermal Units Used by Engine and Feed Pumi*. 

83. Heat of vaporization of steam, 154.6 lbs. absolute, 859.4 B. T. U. 

84. Heat of liquid of steam, 154.6 lbs. absolute, 332.5 B. T. U.' 

85. Heat of liquid, jacket and reheater drain, 328. 3 298.7 B. T. U. 


Percentage of 10.12 lbs. 

i2 Report of Experts on Engine No 3. 


116. Horse power of plungers,.. 599 10 H. P. 

117. Friction horse power, 44 3<3 h. P. 

118. Efficiency of mechanism, 93-12 per cent. 

119. Friction of mechanism 6.88 per cent. 

Boiler Trial. 

Kind of Coal. 
Pittsburg. Pocahontas. 

120. Number of boilers in use, 2 2 

i2r. Duration, hours, 72>o8 ?2 l6y 

Average Pressures. 

122. Steam pressure at boilers by gauge, lbs.,... 140 T40 

123. Atmospheric pressure by barometer, lbs.... 14.59 , 4 .6 2 

124. Absolute steam pressure, lbs., 1S4 . 59 ^Tj^ 

5. Force of draft in inches of water, in., 0.5 0.5 

Average Temperatures. 

2 6. Of escaping gases from boilers, deg., 438.5 457#5 

127. Of feed water before entering heater, deg., 79.0 83. 

128. Of feed water on entering boilers, deg.,... , 4 ;. 2 , 44 .'- 


129. Moist coal consumed, 11 67.917 63,591 

130. *Wood consumedat5opercent. weight lbs., 7?2 2 - 

131. Moisture in coal, per cent., o 7 26 

132. Dry coal consumed with wood equivalent lbs., 67,995 61 692 

133. Total refuse, dry lbs., 2>025 2 '~ 

134. Total refuse, dry, per cent., 2 _ g8 6 , 

135. Total combustible, lbs., 65,970 58,843 

136. Dry coal consumed per hour, lb... g43 ~ '-%- 

137. Combustible consumed per hour, lbs., 9I5 8l - 

138. Calorific value ofonepound of coal by analy- 

3 -> 

«*B.T.U., I3 , 226 

1 4,924 

Quality of Steam 

139. Moisture, 


0055 °-°o55 


1. Pumped into boilers and apparently evapor- 
ated, lbs _ _ 

_____ 0*7,740 593,573 


Specified by contract. 


141. Actually evaporated corrected for quality of 

steam, lbs., - 5 8 4,5°7 59°>3° 8 

142. Factor of evaporation, 1.120 1.118 

143; Equivalent water evaporated into dry steam 

from and at 2 12°, lbs., 654,648 659,964 

144. Equivalent water evaporated into dry steam 

from and at 212 , per hour, lbs., 9 ,082 9,145 

Economic Performance. 

145. Water actually evaporated per pound of dry 

coal, lbs., 8.60 9.57 

1 46. Equivalent water evaporated per pound of dry 

coal from and at 212 , lbs., 9- r M i°-7° 

147. Equivalent total heat derived from a pound 

of dry coal, B. T. U., 9,250 10, 294 

148. Water actually evaporated per lb. of com- 

bustible, lbs, S.86 IO °-S 

149 Equivalent water evaporated per lb. of com- 
bustible from and at 212 , lbs 9-92 i'-- 7 

150. Efficiency of boilers, percent., 7°° 6 9° 

Rate of Combustion. 

151. Dry coal burned per square foot of grate per 

"' hour, lb I2 -7° ir -5° 

Rate of Evaporation. 

152. Water evaporated from and at 2 12 per square 

foot of heating surface per hour, lbs., 2.02 2.04 

Commercial Horse Power. 

153. On basis of 30 lbs. of water per hour evapor- 

ated from ioo° into steam of 70 lbs. gauge 

pressure (=34 V 2 lbs - from and at 2I2 ° } ' 

H. P.,... 26 3- 2 26 5- x 


154. Used per indicated horse power per hour. 11 . 1.47 l 33 

155. Used per pump horse power per hour, lbs.,.. 1.58 1.43 


Report or Kxj'erts on Engine No. 3. 

Results of Coal Analysis. 

l'"i burg, Pot ahonta* 

156. Moisture, [.78 per cent. 0.73 per cent 

157. Carbon, 75.65 per cent. tf?., 1 ,! percent. 

158. Hydrogen, 5.00 per cent. 5.15 percent. 

159. Oxygen, 9.73 per cent, 4.65 percent. 

160. Nitrogen, [.50 per cent 1.25 percent. 

161. Volatile sulphur, 0.64 per cent. 0.76 per cent. 

{(>2 - ^ sl1 yo per cent. 4.15 pei cent. 

100.00 \>> 1 1 -lit. 100.00 per < eni 

Moisture in Coal. 

W< Wcighl Moislure, Perccnl 

bcfon after 02 »g< ,f 

drymg,02 flr-3 02. Mo 

[6 3- F'rst day 4<J2 4HH 40 o ? 

«6 4 . Second day, ,* 2 , t Sj ,. 0.6 

t6 5' l,ll,(| day, , 7< s. 5 , 77 u$ 8 

r66 ' Fourth da y> '«8. 5 184.25 4.25 2.25 

[6 7- I ifth day, 2 , 4 20 * 6 o ; 2 8o 

""■• Si hd *y> 346.75 337.25 9.50 2. 74 

169. Avera moisture in Pittsburg coal, 0i?0 

170. Average moisture in J ihontas coal,.. 2 .6o 

D\ riE and ( !apa< ll v. 

Thecontraci requires a statement of the duty performed during 
«*t« 1 four hours by 100 lbs. of dry coal con umed fo. n,n, , 

delivered into the- reservoir, as del rmined I, weir rremenl 

For this n ison the dutii on this basis, as v I. thei will be 

f :iU ; 1 for ch day. The duties per 1,000,000 heal unit consumed 
tl- engi „ ,iveninth, tables are computed with temfreratun of 

-he steam cylinders taken a. the main feed 1 ,. enters 
lers, and of the jackei n,d reheater drains tak, 1 al th, boil- , 

rh steam used in this computation includes that used bi th< feed 
pump. * 

. , ," l "; ral computation for th, effi, ncy of th, ngin, il , 

r able, p 1; „ the temperature ol rej | the cylinder 

" ' ", al "" P««>P discharge, and o< ,|,. ,„!,,, ;„, 

''' " •" ll "ingtankal thee, ... , ,. , , b „,. 

d pump w i u ii| ( j, ,i. J 































































si ° C 

V3 O fi <U 

a ~ « ^ 


















co : 








r ^ 1 














Report of Experts on Engine No. 3 






s « 














H ^ 






O r^ 







fe On 































C/3* E 








- 2 









~ U? 

r l 












r o 

S V, 

J N 


— sc 











j : 

- -^ 















i— 1 




















O 1 

DC u 

°.D i S op 















^- X 

t- O 









w ^ 

- cc 



r-. 00 


— ' 













HH 1 




— j 




• 63 ~ 


D 1 






2 £ 1 

£. C 



: 10 
: 10 ^0 

00 r^ ri 

^ Tf lo 

H W M 








J . 






• - 

f m 








3 *^ 
1 — 1 t/i ■ 









S - •£ H* ^ 

2S I fen 





^ — 
















> C 

> c 

1 — ^ ^ 

: c 

; C 

> c 

) c 

> c 

> c 



_; ^ "3 I 

: dc 




O — r 



1 r 


< — O k- 



_— ^ 






r ■ 


1 C 








32 £5 



/* O 


- N 





CO ; 



• — 

, M 













T3 T 






^ t: 



» w 










Built fob the Louisville Water Company. 


i , i i<Ai COMPI I \M i FOB El I i Ol I •• ' NE - 

Duration of trial ~ l ** hl IO rnin 

I il i. mnl I revolutioi 160,666.5 

k olutions per minul 574 

\ rap I. II. P ... 4 ° 

Wei pump horse power, 10 

I riction H. P., ... A ° 

1 ri< hum II. P. per cent., 

Efficient ' 1 m -in, \> nt., l - 

Total dry steam u d Ijy ei •••• ■•• I » 1 2 ~< "' 

fatal i« ist si i" u d by engine,. '•' ll) 

1 , ,1 moist I m u 1 by cylinders,. ll> 

Total moisl im used b) nd reh 79S 

•I i moisl I ira .1 dbj k "1 reh. . p 

Dry '.I used per I. H. P. i- hour,. " ll 

Moisl im nli ; 1L r " r n0UT >" ' U 

sed per I. 11 P. per hour 1 ylin • ° lb 

1 up '5' "' 

I >ry steam u 

M in al "i pi 

1 " 

,1 ol li<i l i' tl o( air pump dis< liar 

it of liquid ol 1 eel and rel ' drain ; l 

Mean tempei ure of 1 m gin air pump d 

M ,,1 I mp< iture of 1 "> of ja< k< I tnd 1 3 


n, ,t ol vaporization ol im upply,. 

II, t | of liquid ol am ipply,.. ' l: ' j 

Di 1 ,ni in mixture 1 b> 1 ' ,-• ° 

l; i 1 pei pound of mo ' im used b) 

il . . ^45x86o + :• o 1134 ,; ' l 

I, 1 \ per pound ol m im 1 d l» "«1 

■vl ters, 0.9945* *°- 6 + 

I- 1 1 1 through cylinders in 141! hi 1 07 ' '' 

i: 1 U pa through u ' r n 

1 ,1' hi B " [ 

ML, m Q III 

1: 1 . 1 sing tl h ei '• " 

B r. U. used 1 I H P inute (n I si 1 46 B. I I 

. I I 


1. rmodynamic efficiency >f ■ - 

Ptaneer work perforn I in 1 | },' h ' M ;M 




1 ii • 

1 8 Report of Experts on Engine No. 3. 

*Mean duty per 1,000,000 B. T. U. used by engine 

al o»e, 138,126,000 ft. lbs 

*Mean duty per 1,000 lbs. moist steam used by engine 

alone > 150,838,000 ft. lbs. 

*Mean duty per 1,000 lbs. dry steam used by engine 

atone > 151,672,000 ft. lbs. 

*Based upon plunger work. 

f$j&Orf QCZ*Z£-Q(£ 

Built for the Louisville Water Company 



The water pumped by the engine was forced through two lines of 
30-inch Pipe to the distributing reservoir. At the reservoir the water 
was delivered into a masonry chamber 24 feet long, 12 feet wide, and 
23 feet deep. At one end of this chamber there was a permanent 
weir having an edge of brass .50 of an inch in width, finished per- 
fectly level, with a sharp edge on the upstream side. The weir ex- 
tended entirely across the chamber, but during the trial the width of 
the weir was contracted to 10 feet. 

About 7 feet above the bottom of the chamber a tight floor was 
laid, covering the entire bottom with the exception of an opening 
2x12 feet, next the wall of the chamber, at the end opposite the weir. 

The water entered the chamber under the temporary floor, passed 
up through the 2x12 feet opening at the end, and then passed through 
a rack 10 feet high and 12 feet wide, made of ^-inch strips of wood, 
with i^-inch spaces between, extending from the floor to the top of 

the chamber. 

For the purpose of obtaining the head on the weir, brass plates 8 

inches square, ^ inch thick, having sharp edges, were placed at both 

sides of the channel 18 feet back from the weir and 6 inches below 

the crest. The faces of these plates were smooth, plane surfaces, 

placed parallel with the sides of the chamber and the direction of the 


A hole y of an inch in diameter, in the centre of these plates 

was connected by #-inch pipes, with a cask in which the water level 

was observed by means of a hook gauge. Very careful measurements 

were made to establish the zero of the hook gauge at the level of the 

crest of the weir, and also to test the height of the weir at different 


During the whole trial observations of the depth of water flowing 

over the weir were taken every 10 minutes, and during the first and 

last hours they were taken every minute. The water in the channel 

of approach and while flowing over the weir was carefully protected 

from the effect of the wind. 

The quantity delivered has been calculated by the well known 
Francis formula, as the conditions complied more closely with those 

20 Rei )ri i Experts on E\ im No. 3, 

obtaining during his experiments than with thoa of the Inter experi 
raent of 1 nd 

i i irmul; 

Q= 3 [L— o. i n[(H + h — h~] 1 T H -f h -l,- 1 

( J— 1 ' ubi< u -i id. 

H=T1 , : depth of water on the weir.< In due to il irelocil I ippi 

D==a the numl ml < utractiom 

L=io i the J 'i the weij . 

The iu '-' (J v w r) uni nn, seldom \ ing i tlian .003 ol 

i irii 11 h( 

1 ' • ' pths oi water ai 1 julti quantities d( 

I h il; ir wei 

o p. ,. i, |:ro 1 \ . 00 

'• " t° 1 '• April 0804 106,, 10 

):001 '- ' '' o . , 5 100 

; v P ril *> M 0.84606 1 ,000 

v ■ '■ P.M.April ,6,4 > 

: V P Vpril 0.84 ... ,, ,0 

16,4, S co 

V ' ' M ' ■ u o 

ing the dayi i 

■ I «t. I, 1 the 

p plun 


n " from ti 1 ■ .,,, ,,, a ,* n j 

' 1 ' ■ d 1 

■J to ee that 1 

1 , 

h tl 



• 1 ■ 

' T t 

Built for the Louisville Water Company. 


the faces of the valves and valve seats, and after a careful considera- 
tion of the question, we concluded that the slip, as determined by 
the weir measurement, was probably correct within i or i l / 2 per cent. 

The Boiler Performance. 

The boiler performance was unsatisfactory with either kind of coal. 
While that with the Pittsburg coal was perhaps as good as could be 
expected, considering the calorific value of the coal and its capacity 
for covering the heating surfaces with soot, something better than the 
result obtained was looked for with Pocahontas coal. The latter coal 
was a poor sample of its kind, as shown by the analysis herewith 
given, and an inspection of the coal corroborated this opinion. Never- 
theless, the efficiency of the boilers with Pocahontas coal was lower 
than should be expected with an internally fired boiler of any kind. 
This is explained in several ways, the principal reason being, doubt- 
less, the fact that its use followed that of the Pittsburg coal without 
any cleaning of the boiler except the daily blowing out of tube vith 
a steam jet. Thus the heating surfaces were covered with soot from 
the Pittsburg coal, and the Pocahontas coal suffered in consequence. 
This is borne out by the higher temperature of the escaping gases. 
Another cause of low efficiency with Pocahontas coal is the tart of 
its being the first coal of that kind ever used by the firemen. 


The duties, steam consumption, mechanical and thermodynami< 
efficiencies of Pumping Engine No. 3 < as given in the preceding 
pages, are remarkable in establishing this a. the most economical 
compound or double expansion steam or pumping engine that has 
ever been tested, so far as we are aware. It is furthermore remark- 
able in its mechanical efficiency, as shown by the small friction of the 
machinery, and this contributes considerably to the duty. 

The pump performance, as shown by the pump indicator diagrams, 
is remarkable for freedom from shocks, although there is quite an 
audible shock in the pumps when the discharge valves close. This 
is doubtless due in part to the fact that the valves and seats are wholly 
of metal, and in part to the sounding-board effect of the enormot; 
mass and surface of iron of the pumps and the wooden caisson on 

which they rest. 

The action of the steam valves and valve gear are excellent, and 

the same is true of all parts of the mechanism. 


Report of Experts on Engine No. ^ 

The results given in this report show that all requirements of the 
guarantee of this pumping engine are much more than fulfilled. 

It is with great satisfaction that we are able to make this statement 
without qualification, and to further state that the whole plant is a 
great credit "to the designers and 1-uilders. 

Respectfully submitted, 

Expert for the I. P. Morris Co. 

F. W. DEAN, 
Expert for the Louisville Water Co. 



Leavitt Pumping Engine 





(Member of the Society.) 

Presented at the New York Meeting (December, 1894) of the American 

Society of Mechanical Engineers, and forming part of 

Volume XVI of the Transactions. 

Trial of the Leavitt Pumping Engine. 






(Member of the iety.) 

In April of the present year the writer, as expert for the 
Louisville Water Co., Louisville, Ky., and Mr. Dexter Bracket*, as 
expert for the builders (the I. P. Morris Co., of Philadelphia) of the 
new pumping engine at Louisville, Ky., conducted a contract trial of 
six day's duration. The engine ran 144 hours and 10 minutes without 
a stop, which is the longest test run on record, and established itself as 
the most economical compound engine that has ever been tested, so 
far as the writer knows. The result is phenomenal and is of great in- 
terest at the present time on account of tests of some recent high ex- 
pansion engines with cylinder ratios of 7 to 1, an account of one of 
which the writer gives in another paper. It also has great interest in 
showing how closely reached by this engine are the records of many 
triple expansion engines. The writer believes, however, that a triple 
expansion engine designed on the same lines will lower the steam con- 
sumption by a paying percentage. 

The engine referred to is Pumping Engine No. 3, of the 
Louisville Water Co. It is of the well-known Leavitt type, having 
two vertical inverted cylinders, the piston rod of the high-pressure 
cylinder being connected by links to one end of a beam, and the 

^Presented at the New York meeting (December, 1894) of the American 
Society of Mechanical Engineers, and forming part of Vol. XVI, of the Trans- 

28 Trial 01 the Lea vitt Pvm ping Engini 

low-pressure similarly to the other end of the beam. The main shaft 
is at one end of the engine, and the connecting rod passes from a pin 
in the upper part of the beam to the crank-pin. The steam pistons 
have opposite motions in consequence of this arrangement, and the 
exhausts from the ends of the high-pressure cylinder pass to the cor- 
responding ends of the low-pressure cylinder. There art- two reheating 
receivers between the cylinders composed of small brass tubes, inside 
of which is live steam of boiler pressure, the exhaust steam passing in 
contact with the outsides of the tubes. Both cylinders are steam- 
jacketed on heads and sides with steam of boiler pressure. 

Each steam-cylinder is provided with four gridiron valves operated 
by Leavitt cams. The point of cut-off in the high-pressure cylinder 
is automatically determined by a ball governor, but that of the low- 
pressure cylinder is fixed. The engine is of the most massive 
character, the weight being tar greater than that of any other pumping 
engine of the same capacity. The pumps are located directly under 
the engine, and the plungers are connected to the beam at such points 
that, while the stroke of each steam piston is 10 feet, that of each 
pump plunger is 7 feet. The plungers work vertically and are of the 
differential type, being single acting on the suction, and double acting 
on the discharge. The engine is provided with a surface condenser 
and vertical double acting air pump. 

On account of the rise and fall of the Ohio River the bed plate 
of the engine is placed above the highest high water mark, while the 
bottoms of the pumps are sufficiently low to take water at the lowest 
stages of the river. The distance from the bottoms of the pumps to 
the Dottom of the bed plate is 61 feet. 

The trial consisted of ascertaining the duty by weir measurement 
t the reservoir and nearly or quite all other data of interest That 
part of the trial relating to the engine only will be here described 
The engine is worked by steam of 140 pounds gauge pressure at the 
boilers and this is conducted through l8 o feet of steam pipe, well 
covered to the engine. At the engine the total per cent, of condensa- 
tion in this pipe and priming of the boilers amounted to 2 JU> per 
cent and al of this but t - of , per cent, was thrown out b/ a 
separator The steam pressure at the engine near the high-presLe 
Imder fell to i 37 pounds by gauge. pressure 

At the beginning of the trial the steam pressure in the two boilers 

used wa at about 9 o pounds, and just before starting the engine he 

water-level was marked in both boilers. Immediately after sZjnl 

e engme. 6 days IO minutes later, the same pressured J3S£ 

Triai i i he Li .win Pumpin E ^ine. 


From tl til w ght of -team ei the st< n-pipe ther 

have been deducted tin ' im '1 by the tlorime nd tl, ter 
ii ved by the sepat r. In the appropriate pin the mo 

shown by the calorin r was i hi' /.. v. 

stated in tern of dr. • ira. 


High ! / 

i .1, i w I 

I .-J.— N ii 



Trial of the Leavitt Pumping Engink 

Indicator diagrams were taken every hour throughout the 144 
hours on separate indicators at each end of each cylinder. These 
diagrams were taken with great care, and after the trial the scales of 
the indicator springs were determined at the Brooklyn Navy Yard. 

Fig. 53. — No. 123. Low- Pressure, Top. A = 2.8"> 



— 4 

Fig. 54.— High- Pressure, End. 

Fig. 55.— Low-Pressure, End. 

Before the trial the engine had been run the greater part of a 
jean 1 he and cylinders were thoroughly covered wi h a non- 
conductor of heat. 

The following are the leading dimensions of the engine • 

Trial of the Leavitt Pumping Engine. ri 


Type, Leavilt compound vertical inverted beam fly-wheel. 

Diameter of high-pressure cylinder, hot 27.21 in, 

" low-pressure *' " 54.13 in, 

" fly-wheel .. 36 ft- 

" u high-pressure piston rod ..... 5£ in. 

€i " low-pressure a u *,. 6 in. 

Stroke of each piston.. 10 ft. 

Mean clearance of high-pressure cylinder . 1,585 °| 

ft " i6 low-pressure 4< 1,530 °J 

Diameters of each differential plunger .. 34 in. ; and 24 ,' in. 

Stroke " a " " 7 ft. 

Mean ratio of steam piston areas.. 4,015 to 1. 

Volume displaced by plungers during one revolution of engine.. 660.30 gallons. 

Diameter of each discharge pipe .. 24 in. 

Results of Engine Trial. 

Duration.... 144 hrs. 10 min. 

Total number of revolutions 160,666.5 

Average number of revolutions per minute 1 u 4 

" piston speed p t minute — — 371.48 ft. 

u plunger speed per minute — 260.04 ft. 

Average Temperatures. 

Of Engine room . ■ 00° to 

Of external air.. 48° to 86 

Of main feed at weighing tank 81.2 

Of " ci on entering boiler 108 

Of jacket and reheater drain at boiler 

Of mixture of feed-waters ••• 

Of water in pump well - ♦ -•• 

Average Pressures. 

r o 

Of atmosphere by barometer 

Of steam at boilers by gauge * 

a << « " absolute 

<< " " engine by gauge - — - 

<< « " u absolute 

Of initial steam, high-pressure cylinder, absolute 

Of terminal pressure, low-pressure cylinder, absolute... 
Of back pressure, low-pressure cylinder, absolute 

Vacuum by gauge * - * ** 

Water pressure by mercury column 

Height of mercury zero above water in pump well 

Total water pressure 

Equivalent head 

Steam used by Engine. 

:;•_• ;° 



14.00 lbs. 

140.00 lbs. 

154.60 ll>s. 

137.00 lbs, 

151.60 lbs. 

140.75 lbs. 

7.32 lb 

0.05 lbs. 

27.75 in 

62.50 lbs. 

49.01 ft. 

83.74 lbs. 

J 93.35 ft. 

1,157,923 lbs. 

Moist steam entering steam-pipe ., 

Water drained from separator 23 > 428 lbs - 


Trial of the Leavitt Pumping En<,i\e. 


Steam used by calorimeter... 

Total moist steam used by engine 

srcentage of moisture in steam after leaving separator 

Total dry steam used by engine. 

moist " passing through inner steam cylinders 

tl " u steam-jackets and re- 
heaters im99 

Percentage of moist steam used by jackets and reheaters... 

Moist steam used per hour, per L H. P 

Dry tl li *< ** <* u 

it it it lI 1 1*1 

" by inner clymders. . . 
M «s' " " " " " pump, horse-powei 

J Jry " ft " « a << ti it 

Prevailing point of cut-oft" high-pressure cylinder 

low-pressure ** 

Drop between cylinders 

Compression in high-pressure cylinder 

" low-pressure " 

ilio of ansion by volume 

im accounted for by indicator at high-pressure cut-oil 

in per cent, of 10120 poun 775 Ujs. 

Steam accounted for by indicator at high -pressure release.'. 9.166 II 

" low-] mil . . lu.0.08 II 

" rel. .... 725 II 


IT, Lbs. 
1,133,768 lbs. 


1.127.-:::; lbs. 

943,973 lbs 

1s , .».7 , .*5 11 is. 

12.156 lbs 

|n,] L 'n 11,.. 

I:U25 lbs. 
13.050 lbs, 
0.00 II 

| full. 







\. ,1 e.— The last four it b e r arded ly approximate nly. 

Ave rag r Paw £/ . 

An ' an 1 ctive pr in high-pressun Jm [, ,.. 

Jl ivei loped 1 gh-p sun r 

ii low-j " 

1 '■ ,| ' 1 
1>er in b .-jii ;, u |, 

i% u " sure 

fJ f plungci 


ricti | 

• • ■ 

S.40 i lo 

l 3 lbs 

1 l.i • ■ lb . 
279.00 li. i 
364.40 " 
643.40 •' 

I 16 

i'i i 

19.10 ii I*. 

li • 

■ 12$ 


British Th r! I . EtC. 

n al lm |,r 

ir pumi di 
M reheat* 

■ l mi 



44 jac 


•i liqn 






Trial of the Lkavitt Pumping Engine. 



Dry steam in mixture used by engine 

B, T, U. per lb. of moist steam passing through inner 

cylinders, 0.9945 X S60.6 + 330.90-52.27 = 

B. T. U. per lb, of moist steam passing through jackets 

and reheaters, 0.9945x860.6 + 330.9-306.0 = .. 

B. T. U. passing through cylinders in 144 hrs. 10 min 

B. T. U. passing through jackets and reheaters in 144 hrs., 

10 min 

B, T, U. passing through engine in 144 hrs,, 10 min 

B. T. U. used per I. H. P. per minute (moist steam),. 

Mechanical equivalent of heat (Rowland) 

330 00 = 

222.46 X 778 


li;;t£0 B. T. U, 




Thermodynamic efficiency of engine 

u;:, 171,428 


222.40 " 
778 ft. lbs 


Duties based upon Flunger Work. 

Plunger work performed in 144 hrs., 10 min....\ 171 ,015,o 1 4, ( .n'»0 ft. lbs. 








Duty per 1,000,000 B. T. U. used by engine alone.... 
" " 1,000 lbs. moist steam used by engine alone 
1,000 lbs. dry steam used by engine alone.. 

100 " lt Pittsburg coal 

100 " *' Pocahontas " 

100 '* '•• Pittsburg combustible 

100 " " Pocahontas «* 








Sample indicator diagrams from steam and water cylinders are 
given (Figs. 50-55), and also a combined diagram (Fig. 56). 

This engine is, both in design and results, in striking contrast 
with the Rockwood System engine described in the writer's other 
paper, as shown in the following table : 



Steam pressure absolute 


Ratio of expansion 

Number of revolutions per minute 

Length of stroke 

Piston speed per minute 

Cylinder ratio 

Drop between cylinders 

Dry steam per I. H. P. per hour.. 
Difference in favor of Leavitt 

151.50 lbs.. 
27.75 in. 
10 ft- 
371.5 ft. 
4 to 1. 
12.156 lbs. 
0.684 lbs. =5.3% 


175.50 lbs. 
25.3 in. 

4 ft. 

611,2 ft. 

7 to 1. 

About 14 lbs. 

12.-1 lbs. 

This comparison shows very clearly that the ratio of 7 to 1 does 
not necessarily produce as economical results as a ratio far removed 
from it, even with the additional advantages of 24 pounds more steam 
pressure, 1.6 times as many expansions, four times as many reciproca- 
tions per minute, and twice as great piston speed. It tends to show 


Trial of the Leavitt Pumping Engine. 

Trial of i he Leavitt Pumping Engine, 



that no advantage arises from a drop in pressure between the cylinders, 
if evidence were needed of this. 

It is the writer's opinion that in order to use steam in the most 
economical manner in a multiple expansion engine, the expansion 
must be continuous throughout the series of cylinders (that is to say, 
there should be no drop between the cylinders), and that compression 
should be carried up to the initial pressure in each cylinder. These 
features have been employed to the fullest extent in the Leavitt engine 
which forms the subject of this paper, and the result has surpassed all 
the records for economy of engines of its class 


A test of so much importance as that of the Louisville engine, 
wherein a new record has been established for steam consumption, 
may, with great propriety, be accompanied by some data of the 'og of 

the trial. 

The following are the amounts of feed-water weighed ea< h day ol 

twenty-four hours : 

1st day, 24 hours LV.iJVJ pounds of wal . 

2d » " " 161,799 " 

3,1 " •« " 159,972 " 

I h « « « 161,848 " 

5th " " " L60,438 

6th « 24 hours 10 minutes 164,439 " " 













ea< I 1 

Total weight 

2.55 p. 



i k 













\< { \\ 








13:; 22.5 


52 1 7 





i <. 







329 lbs. 



31,260 rt 


30,989 " 

31,771 " 


31,510 - 



31,876 " 




31,891 M 

1(17 " 

The above unit weight of 52.17 pounds was determined by weigh- 
ing the condensation in a cask of cold water during two hours. 
Thinking that this calibration might be too short to use for a test of 




Trial of the Leavitt Pumping Engine 

six days' duration, a calibration of twenty-four hours' duration was 
made by Mr. Hermany, chief engineer and superintendent of the 
Louisville Water Works, at my request. The engine was run at pre- 
cisely the same speed and against the same head as existed during the 
official trial. Indicator diagrams were taken every hour, worked up 
by me, and gave the same horse-power as on the official trial. During 
these twenty-four hours the condensation was continuously weighed, 
and found to be 31,732^ pounds, which is almost identical with the 
meter results. During this same trial the feed-pump was run by a 
donkey boiler, and its exhaust was not allowed to enter the boiler. 
The weighed feed amounted to 193,133 pounds, which is almost 
identical with the sum of the weighed feed and jacket and re-heater 
returns, as given above, for any single day. 

On October 10 the condensations in the jackets and re-heaters 
were determined by weighing separately, but simultaneously, during 
eight hours, with the following results: 

Average Condensations per Minute, October 10. 

High-pressure Jacket 

7.4938 pounds. 


10.0417 pounds. 

Low-pressure Jacket. 

5.2333 pounds. 

On October 20 the same determinations were repeated for eight 
hours, with the following results: 

Average Condensations per Minute, October 20. 

High-pressure Jacket. 

7,5083 pounds. 


9.9437 pounds. 

Low-pressure Jacket 

5.1417 pounds. 

The engine ran on each of these trials at the following speeds: 

Average number of revolutions per minute, October 10 18.6083, 

" 20 ".'.'.. U.hT,i 

while on the official trial the average number of revolutions per minute 
was 18.574. It is not known what head existed on either October io 
or October 20, and, therefore, what power was being generated 

[NO ,, -This paper received discussion jointly with that by ,he same author on 

"Trials of a Recent Compound Engine with a Cylinder Ratio of 7 to 1 
remarks made ,n debate are published in connection with that paper.] 


and the 

Trials of a Recent Compound Engine. 






Member of the Society.) 

Considerable interest has been recently shown in the perform- 
ances of some compound engines working with high pressure steam ; 
and members will recall a paper presented at the San Francisco meet- 
ing by Messrs. Green and Rockwood, giving an account of trials of 
an engine as a triple expansion engine and, by throwing the interme- 
diate cylinder out of use, as a compound.! The results of the trials, 
which were evidently made with due care, tended to establish equal 
economy of the two types. 

Laying aside for the present consideration of the possibility of 
such results being obtained from well-designed and properly worked 
engines of the two types, the writer desires to give an account of a 
test which he conducted of an engine founded, in its design, upon the 
engine referred to, and embodying what is known as the Rockwood 

This engine was built by the Wheelock Engine Company, of 
Worcester, Mass., for B. B. & R. Knight, of Providence, R. I. and 
located at their mill in Natick, R. I. The engine possessed the cylinder 
ratio of 7 to 1, which, under the system referred to, is held to possess 
special virtue. 

* Presented at the New York meeting (December, 1894) of the American 
Society of Mechanical Engineers, and forming part of Volume XVI of the Trans- 

f Transactions American Society 0/ Mechanical Engineers, Vol. XIII, p. 647, 

No. 499. 


;8 Trials ok a Recent Compound Engine. 

The following are the leading dimensions : 

Diameter high-pressure cylinder, hot 1 *.44 in. 

" low-pressure '* " 48 50 in. 

" high-pressure piston-rod 3.25 in. 

" low-pressure " 4.25 in. 

Stroke of both pistons 48.00 in. 

Mean ratio of piston areas... 7 to 1 

'« high pressure clearance 2 j % 

" low pressure u 2.-'. '< 

The engine is a horizontal cross compound, with the high -pressure 
cylinder jacketed all over, and the low-pressure cylinder on the heads 
only. There was a re-heater between the cylinders. In the writer's 
judgment the jackets were badly piped, and it is doubtful if the jacket 
circulation was good. The re-heater was quite deficient in heating 
surface. The condenser was of the injector type, made by the builder 
of the engine. The vacuum was defective, although very cold water 
was used. 

The engine was four hundred feet from the boiler, which was of 
the Babcock & Wilcox make, but as the pipe and flanges were well 
covered the condensation was not excessive. 

Examination showed the pistons and valves to be tight. 

The feed water was weighed upon correct scales, and was pumped 
by a geared pump. The boiler was entirely separate from others in the 
same plant, and all connected pipes which could carry unaccounted for 
water or steam to or from the plant were disconnected or blanked. 
There were no leaks either in the economizer or boiler, and in the second 
test here described the economizer was not in use. 

In the engine room, indicator diagrams were taken by two indi- 
cators on each cylinder every twenty minutes, the power being very 
uniform. A calorimeter was attached to the main steam-pipe near the 
high-pressure cylinder, and just before it there was located a steam 
separator. The condensation from this separator was kept at a con- 
stant height in a water glass, and the water drawn off was weighed by 
running it into a tank of cold water. The re-heater and jacket conden- 
sations were under control, and were kept at a visible and constant 
height in a glass tube, thus insuring no waste of steam. 

Five different tests were made, but on account of accidental and 
unavoidable wastes of steam in three of them, only two will be quoted 
here. During the two referred to there was a slight leak of steam from 
an expansion joint, and on the last test one safety valve was open three- 
quarters of a minute. These errors are so slight that they can be 

Trims of a Recent Compound Engine. 


The indicator springs were carefully tested by the writer under 
steam, and afterward taken to the Navy Yard at Brooklyn and tested, 
the two results being substantially alike. 

The durations of the tests were shorter than is desirable, but the 
mill hours determined this. 

The folio win st is a brief tabulation of the results: 

Date, 1894. 

Duration of trials 

Average steam pressure near engine 


vacu um 

ratio of expansion by volumes 

number of revolutions per minute.. 

(i piston-speed, feet per minute 

Per cent, of moisture in steam near cylinder 

Total dry steam used , 

Average I. H. P 

Jan. 26, p. m. 

4.'. h. 
159 lbs. 
25.4 in. 




1 .90 ' r 

34.089 lbs. 


Dry steam used per I. H. P. per hour | 12.74 lbs 

Average dry steam used per 1. H. P. per hour... 

Jan- 27, a. m 

5 h. 

158 lbs. 

25.2 in. 



37.077 lbs. 


12.94 lb-. 

12.84 lb-. 

It will be seen that these results show a very economical use of 
steam, and far less than has heretofore been thought possible with com- 
pound engines. If the vacuum had been 28 inches, the steam con- 
sumption might have been as low as 12.36 pounds on January 26, p. m., 




Yig. 57. — Hi^h- Pressure, Head End. 

and 12.60 pounds on January 27, a. m.., if this had not given ris to 
any unfavorable set of thermodynamic conditions. The average of 

these two is 12.48 pounds. 

Sample indicator diagrams are given (Figs. 57-60,) and in the 



Trials of a Recent Compound Engink 

writer's opinion they have a grave defect in showing a considerable 
drop in pressure between the cylinders. The writer is aware that this 
is desired by the designers, but the loss in effect of the steam to which 

Fig. 58.— Low- Pressure, Head End. 

this gives rise cannot be recovered by any subsequent event. More- 
over, this drop exaggerates the difference in the ranges of temperatures 
of the two cylinders, and increases the loss still further by increasing 
the cylinder condensation, according to the well-known and funda- 




Fig. 59.— Hi<rh- Pressure, Crank End 

Fig. 60— Low- Pressure, Crank End 

Trials of a Recent Compound Engine. 


mental theory of the desirability of equal ranges of temperatures. The 
ranges on January 27 were about 144 degrees in the high-pressure and 
82 degrees in the low-pressure cylinders. 

Although the performance of the engine is remarkably good the 
writer believes that it was realized in spite of great defects, and that it 
would have been much better if these alleged defects had not existed. 
The economy, in the writer's judgment, is due to high steam pressure 
with the resultant high degree of expansion, small clearances, and tight 
pistons and valves. 


Prof. R H. Thurston. — The results of short engine trials have 
always been looked upon with much distrust by engineers, when ap- 
parently exhibiting exceptional economy; and the traditional myth of 
the performance of the "record-breaking" Cornish engine of the last 
generation, and of that of the S. S. Thetis, in which, for the time, 
fabulous duties are stated as the results of short duty-trials by famous 
engineers, are a standing admonition to all later experimenters. This 
reproach certainly cannot be urged against this trial, and the profession 
is placed under a real obligation to Mr. Leavitt, Mr. Dean, and Mr. 
Hermany for the admirable example which they have here given of 
deductions based upon unquestionably representative and usual extended 
periods of operation under regular working conditions. The machine 
should, it is fair to assume, be capable of sustaining this duty indefinitely. 
A week's work should be as satisfactorily representative of the capabili- 
ties of the engine as the work of a year. In this case, the result is a 
magnificent one, and designer, builders, and officers in charge of the 
machine have reason for congratulation. I think this "breaks the 
record" for the compound engine to date. A duty of 140,000,000 for 
100 pounds of coal, and of above 150,000,000 for 1,000 pounds of steam, 
represents, probably, not only the best to date for this class of engine, 
but, very closely, the practical limit with saturated steam; and 12 
pounds of steam per I. H. P. per hour seems the limit for pressures of 

125 to 150 pounds. 

The usual conditions of economy are here illustrated fully : dry 
steam, sharp cut-off. full expansion to six or eight pounds absolute 
pressure, free transfer of heat in jackets, with small cylinder-conden- 
sation, no drop between cylinders, and high efficiency of mechanism. 
The jacket-condensation is high, the friction of mechanism extraordi- 
narily low— for such a heavy engine very remarkably so. I doubt if it 


Trials ok a Recent Com rot \i> Engine. 

has ever been equalled, except by the Worth ington class of direct-act- 

ing machines. 



Steam pressure absolute 


Ratio of expansion 

Number of revolutions per minute... 

Length of stroke 

Piston speed per minute 

Cylinder ratio 

Drop between cylinders 

Dry steam per I. 11. P. per hour 

Difference in favor of Leavitt 0.6 > 4 lbs. 

" " " Triple 0.478 lbs. 

151. IK) lbs. 
27.75 ins. 
10 ft. 
371.5 ft. 
4 to 1. 

12.156 lbs. 





175.50 lbs. 

2-V> ins. 


:<■ a 

4 It. 

611.2 ft. 
7 to 1. 
About 1411 
12.84 lbs. 

1.16 = 9 


135.45 lbs. 
27. ti ins. 


5 ft. 

203 ft . 

1, 3, 7 

"i 1.678 ibs 

As presenting an interesting comparison, I have taken the liberty 
of adding to Mr. Dean's table the figures for the Milwaukee engine, in 
order to bring especially the relation of compound to triple, as a com- 
parison of the best work in each class permits now, with a conclusive- 
ness never before allowed. In the collection of data thus assembled, 
we find the triple expansion machine with lowest steam pressure, lowest 
piston speed, and lowest ratio of total expansion, gives four per 
cent, higher economy than the compound, nine per cent, better than 
the hermaphrodite machine, and this means, no doubt, that Mr. Dean's 
statement is perfectly correct; that the triple engine would have proved 
in the hands of Mr. Leavitt as remarkable in its class as is this com- 
pound in its field. The observed difference would be exaggerated, 
were the triple given the advantage of equally high steam and high 
piston speed, and it would seem probable that, under the conditions 
here indicated, the gain by the addition of the third cylinder would 
be something over five per cent. The loss by leaving it out, and sub- 
stituting a receiver with free expansion, would seem, under similar con- 
ditions, to be likely to prove in excess of ten per cent.; a high price to 
pay for the saving of even a steam cylinder with its valve-gearing. 
Mr. Leavitt's success is one iii which the whole profession may find 
cause for pride. 

Mr. F. H. Ball. — This paper institutes certain comparisons be- 
tween the Leavitt Pumping Engine at the Louisville Waterworks, and 
another engine which is described as the "Rockwood System," and 

Trials of a Recent Compound Engine. 43 

certain deductions are made by the author as a result of these compari- 
sons. Unfortunately for some of us, at least, who are interested in this 
subject, we have not been informed as to exactly what the "Rockwood 
System" is. We have had several very interesting reports from Mr. 
Rockwood, of trials of compound engines, where cylinder ratios larger 
than usual were used, and many of us, who believe he is on the right 
track have hoped that he would elaborate his theory in a paper for 
presentation to this Society. If the ratios commonly used are wrong, 
there must be some theory to demonstrate this fact, and to point to 
some other ratio as being better. Mr. Rockwood has told us, on 
different occasions, of his engines, with various cylinder ratios, ranging 
as high in one case, I believe, as 9 to 1. Does his system then consist 
of simply making cylinder ratios greater than heretofore, and does it 
cover all cases from the conventional ratio to infinity, or is there a 
choice in this matter? Mr. Dean seems to think that he has located 
Mr. Rockwood at 7 to 1. Let us proceed on this assumption. 

Referring to the performance of the two engines under considera- 
tion, it must be admitted that the results obtained in both cases are 
phenomenal. Here are two compound engines showing an economy 
that has seldom been equalled by the best triple-expansion engines, 
and never exceeded by them but by a very small amount. The 
Leavitt engine stands at the head, with its 12.15 pounds of water per 
horse-power per hour, and the Rockwood engine a good second with 
12.84 pounds. In comparing these remarkable engines, Mr. Dean has 
made some sweeping conclusions, that perhaps may be fairly questioned. 

On the last page of his paper he uses the following language : 

"It tends to show that no advantage arises from a drop in prepare between 
the cylinders, if evidence were needed of this." 

Also, in the closing paragraph, Mr. Dean says: 

"It is the writer's opinion that in order to use steam in the most economical 
manner in a multiple expansion engine, the expansion must be continuous through- 
out the series of cylinders (that is b ay, there should be no drop between the 
cylinders,) and that compression should be carried up to initial pressure in each 

I must take issue squarely with Mr. Dean, both in regard to this 
being a reasonable conclusion from the figures of his test, and also in 

regard to its being true. 

Taking the question of compression first, where is there, in the 
reported data of these engine trials, one iota of evidence on the sub- 
ject of compression ? Here we have two engines, with widely differing 

44 Trials of a Recent Compound Engine. 

cylinder ratios, tested under conditions that are di>-imilar in almoT 
every respect. In comparing the two engines, the least conspicuous 
difference is in regard to their relative rates of compression. There- 
fore I don't think Mr. Dean is warranted in arriving at any conclusions 
whatever regarding compression, from the figures of these trials. If his 
compression theory rests on any other evidence, I hope he will give it 
to this society in connection with this paper. As against his theorv 
we have the engine trials conducted by Professor Jacobus, reported at 
the Montreal meeting of our Society, in which trials all the conditions 
remained practically constant except compression, and the evidence 
obtained is conclusive that full compression did not in this case give 
the best economy. Does Mr. Dean question the accuracy of the data 
reported by Professor Jacobus, or if not, how does he make his theory 
fit these facts? 

Coming back to the other part of his opinion, he tells us that 
"There should be no drop between the cylinders." Persumably this 
opinion is confirmed in his mind by a study of the data obtained in 
his trial of the two engines under consideration. Let us see how logi- 
cal this looks. 

First, the great dissimilarity of conditions governing these test- 
would seem to make it very difficult to estimate the effect of any one 
of the features of difference, because all of these differences were 
present continuously during the tests, and each producing its own 
modification of the result. 

Second, let us suppose, however, that the case was different, and 
that the two engines were exactly alike in every respect except as to the 
cylinder ratios, and the consequent terminal drop. Let it be assumed 
also that the conditions of the test were identically the same with 
both engines. The Leavitt engine, Mr. Dean tells us, represents his 
theory to the fullest extent. This engine has a cylinder ratio of 4} 
and, without appreciable drop between the cylinders, maintains a 
practically continuous expansion to about 20 volumes. 

The Rockwood engine has a cylinder ratio of 7 and a consider- 
able terminal drop between the cylinders, and expansion is carried to 
about 33 volumes. Between these wide extremes there is a vast unex- 
plored wilderness, so far as any information from these tests is con- 
cerned. If the economy of these engines was represented graphically 
with relat.on to the economy of similar engines with greater cylinder 
ratios than Rockwood, and less than Leavitt, the result would be a 
curve on which Rockwood and Leavitt would appear near that part of 
the curve representing the best economy, and beyond Rockwood at 
one end, and Leavitt at the other, the curve would turn toward a 

Trials of a Recent Compound Engine. 


greater consumption of steam. Suppose Mr. Dean has established two 
points on this curve with the data from these engines. How can he, 
without a third point, locate the curve, and say that Leavitt is at the 
lowest point ? He may with propriety say that this engine shows the 
best recorded performance, and that it is better than the performance 
of the Rockwood engine which he tested, but it seems to me that he 
has no warrant from these figures for saying that "There should be no 
drop between the cylinders," nor that " compression should be carried 
up to initial pressure in both cylinders," because it is only a surmise 
on his part that a still better result than either would not have been 
obtained with t-ome compromise between the two. 

The net result of any engine trial is simply the combined result 
of a great variety of conditions, and hence the uncertainty of attribu- 
ting a good result or a bad one to any one of these numerous condi- 
tions, without having carefully tested for that condition. Anything 
short of this is mere guess work, which we are all privileged to engage 
in as a diversion, but which has little value from a scientific standpoint. 
Mr. Dean finds a slightly better result with the Leavitt engine than the 
Rockwood, and guesses that it is due to full compression and no drop 
between the cylinders. From the standpoint of his theory he finds an 
unexpectedly good result from the Rockwood engine, and guesses 
again, "That it was realized in spite of great defects." Following 
Mr. Dean's example I am inclined to guess that the economy of the 
Leavitt engine is realized " in spite of great defects," and these defects 
I should call the full compression and lack of drop between the 
cylinders, which are the very features he commends as being the full 
realization of his theory. In this matter of guessing we are both now 
on record, and can await the verdict of future experiments. The 
Jacobus tests, already referred to, seem to be good evidence on the 
subject of compression, and if Mr. Dean lias anything else in this line, 
he will no doubt offer it in closing the discussion of his paper. 

On the question of terminal drop, my reasons for differing with 
Mr. Dean will be found on the following pages, which I shall be glad 
to have criticised and discussed by Mr. Dean, or any member of the 


First, assuming that, in a given compound engine, the most 
economical range of temperature and pressure for each cylinder is 
known \ then is it not the function of each cylinder to effect the most 
economical use of steam between the extremes of pressure through 

which it works? 

Second, considering the low-pressure cylinder alone, and assuming 
that a fixed receiver pressure is practically maintained, may not the 



Trials of a Recent Compound Engine. 

economy of the low-pressure cylinder be studied apart from the high- 
pressure cylinder, and is it not true that the economy or wastefulness 
of the low-pressure cylinder cannot affect in any manner the economy 
of the high-pressure cylinder under the assumed conditions as to con- 
stant receiver pressure? 

Third, referring to the high-pressure cylinder, and still assuming 
a practically constant receiver pressure as before, is it not true that the 
economy or wastefulness of this cylinder produces no effect on the 
economy of the low-pressure cylinder, provided the low-pressure cylin- 
der is made to account only for the steam delivered to it from the re- 
ceiver ? 

Fourth, assuming that the foregoing questions have been answered 
in the affimative, let it lurther be assumed, for reasons that will appear 
later, that the boiler pressure is such that a receiver pressure equal to 
the atmospheric pressure has been found the most economical. Under 
the foregoing conditions, then, the high-pressure cylinder would per- 
form exactly the functions of the single cylinder of a simple engine 
without the condenser, because it would receive steam at boiler 
pressure and reject it at atmospheric pressure. 

This brings the subject down to a point where the writer is glad 
to agree heartily with Mr. Dean in his statement regarding the high- 
pressure cylinder, when he says that any loss in effect of the steam in 
this cylinder "cannot be recovered by any subsequent event." If 
this is true, then, for the best results from this engine, it is necessary 
that the high-pressure cylinder should develop the highest possible 
economy when receiving steam at boiler pressure and discharging it at 
atmospheric pressure, and, as has already been stated, this is exactly 
the function of the simple non-condensing cylinder; therefore the 
data obtained in trials of simple engines may be safely applied to the 
high-pressure cylinder of a compound engine such as we have under 
consideration. This opens for us a vast field of research among 
reliable records of simple engine trials, and if Mr. Dean will point to 
a single case where the best economy from a simple engine was 
obtained by expanding to atmospheric pressure, and thus eliminating 
terminal drop, it will greatly fortify his theory. Is it not true, that in 
every instance where simple engines have been tested at various points 
of cut-off, that the highest economy has always been found when the 
expansion curve terminated at a point higher than the atmospheric 
pressure? This terminal drop results in a loss of work, it is true, and 
this loss increases rapidly with increase of drop, as was illustrated in a 
paper which the writer presented to this Society at the Montreal meet- 
ing; but, up to a certain point, this loss is more than overcome by the 

Tri \ls of a Recent Compound Engine. 


resulting increase of mean effective pressure relatively to the cylinder 
condensation. Terminal drop or free expansion develops heat by in- 
ternal work in the steam, and thus produces a superheating effect in 
the steam discharged under these conditions. In the case of a simple 
engine, this superheating is lost by being discharged into the atmos- 
phere, while, with the compound engine which we are considering, 
the low-pressure cylinder utilizes this superheat, and therefore a greater 
terminal drop is permissible than when the cylinder discharges into 
the atmosphere. For the purpose of utilizing the data obtained from 
trials of simple engines in this investigation, a receiver pressure equal 
to the atmosphere was chosen. Whatever can be shown to be true 
with the boiler pressure and receiver pressure, we have assumed will 
also be true with regard to other pressures, to some degree, at least. 
The foregoing course of reasoning is conclusive to my mind that Mr. 
Dean's theory is wrong, and it is to be hoped that this question may 
be definitely settled soon, by carefully conducted experiments, having 

that object solely in view. 

Mr. George I. Rockw wd. — The two papers presented by Mr. Dean 
naturally interest me very much, and I trust I may be pardoned if I 
discuss them at some length; as, though terse (and, I may add, refresh- 
ingly so,) yet they bear with force upon not only the relative thermo- 
dynamic merits of the two engines whose economic performance they 
describe, but also upon the general theory of the high-duty steam- 

Let us refer to the contrast said to exist between these two engines. 

Take, as the first consideration, the steam end of the Louisville engine. 
This may be reasonably regarded as embodying the best design, and, 
perhaps, the best mechanical execution that we can hope to secure in 
an engine having two cylinders of a volume ratio of i to 4, working 
under a steam pressure of 140 pounds, and under pumping-engine 
(that is, the best) conditions. These conclusions are confirmed by the 
news in Mr. Dean's paper of its actual performance; an inspection of 
the indicator diagrams shows that the thermodynamic conditions of 

its operation can hardly be improved. 

Consider, second, the Natick compound engine, which embodies 
in its design the extreme cylinder volume ratio of 1 to 7; it has small 
clearances and large ports in the cylinders, its pistons and valves are 
reasonably tight, though manifestly not perfectly so, as I will presently 
show. It has a relatively large intermediate receiver (a very important 
adjunct to the engine,) which, as Mr. Dean says, contains rather too 
few brass tubes to produce the best steam-jacket effect, although baffle 
plates are used to get the utmost possible contact of steam with tubes. 

4 s Trials hi \ Ri'iM Comi'hixd Engine. 

In one important point the design of the engine is not on "all fours" 
with that of the Louisvilte engine ; namely, it has no barrel jacket on 
the low-pressure cylinder. 

Now I do not agree with Mr. Dean that the conditions of opera- 
tion of each engine are such as to make the comparison of duties 
actually attained a perfectly fair 'one from which to judge between the 
relative economic advantages of the two different systems of designing, 
which, as machines, no doubt these engines illustrate very well. How- 
ever, a pretty fair estimate can be formed if only correct inferences 
are drawn from the data Mr. Dean gives us. Allow me to say here 
that although the different parts of the Natick engine, such as the 
details of the cylinders, the details of the valves and valve-gears, and 
the running parts, and the volume of the receiver of this engine, were 
decided upon by myself, yet I never saw the engine but twice in my 
life: once, after it was erected and had been running some months, 
and once after it was tested. The details of its application to the 
place where it now is I had nothing to do with. 

I think, with the author, that the jacket circulation of this engine 
is perhaps poor ; that the re-heater does its work under adverse con- 
ditions * that the vacuum was not so good, by an amount which I 
estimate from the papers at 1.5 pounds, as in the case of the Louisville 
engine trial ; that the large steam pipe from the Babcock & Wilcox 
boilers — extending out of doors for hundreds of feet — leaked more or 
less at the flange joints. But all the conditions enumerated are adverse 
to the best results by this engine as compared with the pumping-engine. 
On the other hand, it is urged that this engine runs at nearly twice the 
piston speed of the Louisville engine. This point has hitherto been 
considered of much theoretical advantage. I question it, however, 
especially in view of the many recent tests of slow-speed steam jack- 
eted engines in which the economy seemed really improved by reason 
of that slow speed. The larger sizes of the cylinders of the Loui>ville 
engine should more than compensate for any fancied advantage to the 
Matick engine, due to its faster reciprocations. 

The Natick engine had at cut-off in high-pressure cylinder 20 
ounds more steam pressure to its credit than the Louisville engine, 
and perhaps this is a fair point to raise as a disadvantage put upon the 
Louisville engine, though I believe that engine would have done no 
better with the extra 20 pounds than it did do, owing to too small a 
low-pressure cylinder. I 

Now for an estimate of the real advantages of either system over 
the other, as revealed by Mr. Dean's tests. 

First, he makes out an apparent advantage in favor of the Louis- 

Trials of a Recent Compound Engine 


ville engine of 5.3 per cent. I ask, is this figure to be taken as repre- 
senting the true comparative economies of the two types of compound 
engine? I believe it is not, and for the following reasons, partly 

specific and partly general. 

At the trial of each engine the M. E. P. referred to the low- 
pressure cylinder, and the degree of vacuum was: Louisville engine, 
24.9 pounds M. E. P., and 13.4 1 ^ds vacuum: Natick engine, 
17.46 pound, M. E I and 11. 9 pounds va« mm. If the loud on the 
Natick engine could have been enough more to have made u 
vacuum of «3-4 pounds instead of on ir.92 pound I tins 

deer in back pn sure of 1.5 pounds could have been effected i \ 
, , wlded to the M. E. P. of 17- !'< pounds, as is entirely p iible d 
as we should not do on paper, if the proper effect of the better vac im 
on the economy of the Natick engin to be understood, thru 

(1.5 : 17.46 = 8.6 percent.) 8.6 per cent, hum work dom by 
I274 p n ds oi im would immedi ly 1 It. The quantity 12 74 
pounds is now 108.6 per en t of the amount 1 iry I se- 

power of work, io 100 per 1 I would be 1 74 j 10 1QO = 

11.75 pounds steam as the tru mparative • >my of tl k 

engine, as against 12.16 pounds, thai of the Lou ille 
difference in favor of the Natick engine oi ,5 | 1 will not 

try to estimate the harmful efl 1 on the Nati< k ei dut) ol poorly 

piped ja< li 3, insufficient brass tube area in the receb jacket, 
error in dt rmination of its actual performam due to leal 
gteam from th main steam supply-pipe, although il in!) 

something, and perhaps But I should like to 1 

that th ere « 1 1 lIc by the m valve on the t nk end 0! the high 
pressure cylin r, shown | 1 1 > irl) am in Fig. j, | 

4, of the paper on Natick ngine. 

The point of cut-off shown on this card 1 learan. 
at 19 of the stroke, The point of cut-off shown by tl rd t 

22 \ per cent, of the stroke. One would ct to find a lov rmi 
na i p r essu on tl, ard having tl irlier cut-off, instead f whi h tht 
card showing the fewest expansions gives the low ;rminal | 
I estimate then- in pressure due i<> le '■ I n at 1 

pounds Thei appears to ha been a I «n on the other 

strok( », though much 1 >, is I estii the ri in pr< 
leakage to be b as 1.5 poun- W the Nat 

suffe] alo , ficien byre >n of leaky val 3 not 1 

cannot be corn tl] timated. Thus] >wn that, il th t 

f a ii th d idvant ges were to be r, tl 

steam "l- - to &™ r of the Nati< * W ° M ^^ 


Trials of a Recent Compound Engine 

> ) 

> t 

larger than 3.5 percent. I believe I have thereby shown that these 
data also reveal the engines of the style of the Natick compound as 
better than ordinary compounds. 

Mr. Dean touches upon the theory of the successful operation of 
the Natick engine in these words: " * * ■ the economy of the 
Natick engine is due to high steam pressure with the resultant high 
degree of expansion, small clearances, and tight pistons and valves, 
He might have added, "and to the relatively very large port areas, 
as there is probably no other kind of engine extant having so little 
clearance space. 

Mr. Dean also says, "Although the performance of the engine is 
remarkably good, the writer belitves that it was realized in spite of 
great defects," but how does Mr. Dean harmonize these apparently 
conflicting ideas? If this engine does remarkably well; in spite of 
grave defects ; then let us study somewhat the nature of the alleged 
defects, to find out if such they really are. 

To define the Natick engine as simply as possible, it is a triple- 
expansion engine with the intermediate cylinder omitted, and with an 
intermediate receiver substituted therefor. 

The notion that the only effect of an enlarged intermediate reser- 
voir between the first and third cylinders is to drain water out of the 
incoming steam and to heat the steam (in case a steam-jacket is used) 
is one that appears to have taken root in some minds, and I would 
like now to uproot it. That I may explain clearly what I mean, allow 
me to refer you to the combined diagram of the Louisville engine on 
page 8 of Mr. Dean's paper. It may be noted there that no drop 
occurs at the terminal of the high-pressure card. But what happens 
on the return stroke? The pressure falls rapidly to a point about in 
the centre of the back pressure, at least eleven pounds lower than the 
terminal pressure of the high-pressure diagram. Is this to be classed 
as "drop" or not? and does it increase the total range in temperature 
in the high-pressure cylinder? While the bugbear, "drop," is vari- 
ously defined, still, as it brings with it all the disadvantages of drop, 
in my view, it is "drop;" it does tend to increase the temperature 
range in both cylinders. 

Now, we read the receiver volume was about seven-eighths of the 
high-pressure cylinder volume. What would be the effect on the back- 
pressure line of the high-pressure cylinder diagram if, instead, this 
volume were, say, three times or more the volume of the first cylinder? 
Would not the effect be to cause nearly all the "drops" to take place 
at the terminal of the high-pressure card? It would cause a nearly 
straight back-pressure line in high -pressure cylinder, at a pressure equal 

Trials of a Recent Compound Engine. 


to the lowest pressure now occurring in the high-pressure cylinder. 
This would give no greater temperature range in the first cylinder, but 
it would, on the other hand, considerably reduce the range in the sec- 
ond cylinder. Not a pound of pressure would be sacrificed at cut-off 
in second cylinder, and the work done by the engine would be slightly 
increased, although, theoretically, there would be a slight loss of area 
at the toe of the high-pressure card of the combined diagram. 

I ask, would it not be a good thing to do to lower the initial pres- 
sure and temperature in the low-pressure cylinder if unaccompanied by 
any corresponding increase in temperature range in the high -pressure 
cylinder? But all this would be the result of increasing the size of the 
intermediate receiver, and it can be obtained in no other way. The 
mechanical advantage of not striking so heavy a blow on the large low- 
pressure piston is also considerable, though apart from the phase of the 
question which I would like to present. 

Now, in the test of the Natick engine the receiver pressure was 
carried relatively higher than I would desire it to be, owing to the fact 
that it was somewhat underloaded ; but still the receiver volume is 
nearly or quite as large as that of the low-pressure cylinder, and so it 
has the effect of decreasing uniformly the back pressure on the first 
cylinder, in this case fourteen pounds. Thus it makes the range in 
temperature in the large cylinder also much less, and-please mark this 
statement— thereby contributes to the economy of the engine as a 
whole How does it do this? Let this question be answered by a 
consideration of the grounds upon which the -well-known and funda- 
mental theory of the desirability of equal ranges of temperatures 

rC5tS This theory asserts that in each of the cylinders of a compound 
engine an equal amount of cylinder condensation will occur, provided 
that the range in temperature in each is equal. Could anything be 
more erroneous on the face of it than that proposition ? What account 
does it take of the fact that the low-pressure cylinder of, say, the 
Louisville engine has four times the exposed area on its piston and 
cylinder-head faces that the high-pressure cylinder has? A moment s 
consideration should show that a unit of area in either cylinder ex- 
nosed to a degree difference in temperature will, of er conditions bet H 
Uenticak condense an equal amount of steam, unless, indeed, there 
be some at present unknown dynamic influence upon the incoming 
.team tending to augment condensation. 

Thus it seems to "stand to reason" that in the case of the 
Natick engine, if the ranges in temperature were maintained equal 
in each cylinder, with a difference in piston areas of 7 to t, there 

52 Trials of a Recent Compound Engine. 

would constantly be many times the condensation occurring in the 
first cylinder occurring all the time in the second cylinder. 

It appears to me plain that the maximum efficiency of the entire 
engine is reached when there is an equality, not of temperature 
ranges, but of amounts of cylinder condensation I the condensation 
occurring in the first cylinder being just sufficient to, after re-evapor- 
tion at exhaust, take the place of the condensation bound to occur in 
the succeeding cylinder. Thus, as Dr. Thurston has well said, the most 
wasteful cylinder in series is the measure of the loss from cylinder con- 
densation ; plainly, we can do no better than to make each cylinder 
equally wasteful by adjusting the range in temperature in each cylinder 
so as to produce this result. 

There is but one way to secure an equality of condensation in 
the two cylinders of the Natick engine, and that is, as I have just 
attempted to show, by employing a very large intermediate receiver. 

This will of necessity produce some drop at the terminal of the 
high-pressure card, whereas the intermediate cylinder would prevent it 
utterly. To the extent that "drop" is a net loss the use of an inter- 
mediate cylinder would be a gain, for I realize fully that part of this 
loss 'cannot be recovered by any subsequent event." But after we 
have admitted this fact we are still no wiser than before; we must 
arrive at some idea of the net extent of the loss by "drop," that is, 
the net loss "after all the bills are paid," to use a business man's 
simile, and then, if we find it to be serious — say, something over two 
or three per cent. — we can make use of the intermediate cylinder. 

Now, it is apparent from an inspection of the combined diagrams 
referred to on page 8) that one loss due to making use of an interme- 
diate cylinder is the loss due to wire-drawing in getting the steam out 
of the first cylinder and into the second. This of itself is a greater 
loss than the triangular area lost through ordering the point of cut-off 
at a point on the entire expansion curve that is lower than the termi- 
nal pressure in high -pressure cylinder — a fact which is the cause of the 
drop in the Natick engine. Then a certain, and relatively considerable, 
portion of the toe of the high-pressure diagram is so much lost work, 
owing to the fact that it is too little to overcome the friction of the 
engine, as has been pointed out by Professor Gale. 

My belief is that when such practical considerations as those just 
given are arrayed on the credit side of "drop"— and, be it under- 
stood, I here allude to a small degree- of drop, say not over 30 pounds 
—such as we can get along with where the expansions are so many as 
in the Natick engine, the preponderance of power felt at the piston- 
rod will be found to be in favor of the two-cylinder rather than the 

Trials of a Recent Compound Engine. 


three-cylinder engine, where pressures of 160, 170, or 180 pounds are 

to be obtained. 

I have dealt with the idea that the office of the receiver in the 
Natick type of compound engine is to produce drop in pressure at the 
terminal of the high-pressure cylinder stroke ; that there is practically 
no loss from "drop" in that engine; and that in any compound 
engine it is necessary to sustain, not an equality of temperature- 
ranges in the two cylinders, but an equality of condensations. I would 
now like to look at the question in another light, and will try to .how 
that, leaving the low pressure cylinder quite out of the account, there 
is still no greater loss from cylinder condensation in the Natick 
engine, even though the intermediate cylinder is not employed, than 

would be the case were it in use. 

Suppose the engine to have an intermediate cylinder of a 

diameter of, say, thirty or thirty-two inches; that is, give the ervgini 

what would be a standard intermediate < ylinder. 

Suppose the three points of cut-off to be so adjusted as t< ;iv< 

equal ranges of temperature in each cylinder. We would then have 

the kind of practice desired by Mr. Dean. 

The relative areas of the high and intermediate cylinders ire ti 

each other as 1 to 3, and the ran 3 in temperature are presupposed 


Now, it seems to me that, in order to prove that the intermedial 

cylinder is an "ameliorator" of the loss in the entire engine due to 
cylinder condensation, it must first be shown that I linder conden- 

sation, by a considerable amount, gets by the intermediate piston 
without doing work in that cylinder as st m than would ape from 
the high-pressure cylinder, were it to be subjected to twi- the range 
in temperature happening when both cylinders are in use, by the 
instrumentality of an enlarged receiver. Perhaps it is unnecessary to 
take time to show that the effect of either the intermediate cylinder or 
of the laro-e receiver upon the conditions under which the low-pr- are 
cylinder takes its steam is identical in either case. 10 that, as I have 
said, that cylinder may be left out of a< count, in calculating the dele- 
terious effects on the economy of the engine by reason of leaving out 
the intermediate cylinder 

The question, therefore, is: "Does more condensation and 
re-evaporation take place in the high-pressure cylinder- 
the temperature-range and one-third the area of the intermediate 
cylinder— than takes place in the intermediate cylinder, if used ? " 

To ask this question is also to answer it, I think, in the negative,, 
in the light of what lias been said above. 

having twice 


54 Trials of a Recent Compound Engine. 

To return to the author's indictment, that the Natick engine 
labors under great defects ; I have mentioned that many of the defects, 
such as defective jacket circulation and defects of that order, are of 
themselves a sufficient cause of the difference in economy actually 
observed between the two engines; I agree fully with him in the 
abstract proposition that the highest economy to be realized in the 
perfect engine — that is, in one having non-condensing cylinder sur- 
faces and frictionless parts— is to result from the combined influence of 
two conditions — using a volume of steam at the highest possible pres- 
sure, expanded the utmost number of times. 

The Natick engine is the embodiment of this principle, so far as 
the principle can be embodied. It uses steam of a higher presume 
Than does any other compound mill engine of which I have any know- 

It expands a volume of it sixi\ per cent, more times than tin- 
ratio of expansion in the Leavitt engine — the ratio being i to 33 for 
the Natick engine and 1 to 21 for the Leavitt engine. If the Leavitl 
low-pressure cylinder had been fifty per cent, larger it would havi 
bled the expansions to be on a par with the ratio in tin Natick 
in< but the increase in economy would only result, I ventm to 

predict, if an enlarged r -real D used. 

But > to the size of the recei .nd the volume-ratio in the 
•Jatick engine being great defects in its design, 1 confess, for reasons 

tated, 1 < ini it quite t in that light, and mistrusi I shall 

1 able to it 1, unless I am given more information of a kind con- 
trary to that now in m\ on. 

There is one other minor nd last pe< I of the matter that I 

u it bring briefly to your attention by quoting the ti ig 

"On lIIow doi not make a summer." This Loui ille engin< m 

not only brol il ] >us 1 ords, it 1 I them oul oi sight they 

an not even in th< at all. 

Note the p brm of the Pawi I I compound pumping 

th. of the g Li U lis tandem < mpound at th< Plymouth 

\\ ,, — fifteen or twenty | cent. l< • momical, though 

i, fully j.^ l »(1 conditioi N< thai of the tri] , 1 

Laketon pu pii worl ig with steam at on hundred and ft It) 

nd 1 per cent. l< mi< al, Note thai of tht 

Euro] tripl pa > mill me, thi er Corli • f lai \u 

d ign, it distant I by this Loui pound r 

■inj I < 1 i the 1 foi nam oi th 

— ii i u] 1. .1 it dou U - < - .nit] I'- 

ll ) ' u 1 to warrant an impea< h 


Trials of a Recent Compound Engine. 


ment of its design, especially when four other engines of the same type 
have all given equally good or better accounts of themselves ; whereas 
we cannot, with certainty, get a plain compound Corliss mill engine to 
do as well as fourteen pounds, try as we will. 

Dr Emery.- -It is known by many present that several of the 
problems under discussion were examined by me about twenty years 
ago The lessons then learned have not lost their force in many re- 
spects The later engineers have had an opportunity of experiment- 
ing with higher steam pressure and more perfect mechanism, and have 
obtained much more economical results ; but it is a question .f such re- 
sults are not due entirely to these two features. I class reduced clear- 
ance with more perfect mechanism, for the reason that the mechanical 
details of the engines were substantially the same then as now. lhere 
is a tendency, however, to theorize, as to feature, other than those 
mentioned, and we are fast reaching a condit.on of ultra theory and 
ultra expansion, like that developed for the older tj of engmes dur- 
ing the war, when the Winooski and Algonquin ran their celebrated 
dock race here in the city of New York. It will to ted tha 

on the last-named vessel , S to 2 o expansions were attempted u nngte 
cylinder with 80 pounds steam pressure, while, m the other , de- 

signed by Mr. Isherwood, 45 P°»"ds steam p, - u I, cut-off 

at bout A of the stroke, but with a valve m. . «ly that the 
v r ual cut-off was at about A. The low-pr ure steam machinery 
"Sed more steadily than the other, used less steam per horse-pow r, 
and did not break down, where., that using the high-, d d. 

Til showed that there was more to the que on than mere :ory 
I L case .lie expansion was carried to an extreme unwarranted b 3 
e condi on: so that the more simple machiner with iess ton 

ha n was warranted, gave the better results. History -P<«£ 

previous practice. In discussing i cylinder, and made the 

tn ,hp verv low mean pressure in the large cylinder, 

to the very iow t ; n ^rmediite cvl nder could hav 

56 Trials of a Recent Compound Engine. 

designed operate with as low a mean pressure in the low-pressure cylin- 
der as I had criticised. We ought by this time to know all about the 
results with low-pressure steam, as very many experiments have been 
made with it. Mr. Isherwood's books are full of such experiments. 
Those made on the Michigan, at Erie, Pa., settled many of the questions, 
though others are equally applicable, more particularly those with which 
the speaker was connected, known as the " Novelty Iron Works Ex- 
periments," of which a table has been published, without explanations, 
in Appleiojfs Cyclopedia of Mechanics, and Vol. II, American edition, 
of Weisbacti s Mechanics. The general result is well-known. Engines 
using 15 pounds of steam were more economical than those using 5 to 
10 pounds; 25 pounds pressure was found still more economical, and 
40 pounds more economical yet. The last-named pressure is at present 
out of the question for the large cylinder of a compound engine. In 
fact, there would be some gain by compounding with such pressure, but 
in regard to using steam at a pressure below that of the atmosphere, 
and at 10 or 15 pounds above it, there is no question whatever; the 
latter is much more economical. The terminal pressure in a low-pres- 
sure cylinder should be high enough to insure thorough drainage by the 
sudden expansion during the exhaust, the gain in this way being greater 
than the loss caused by reducing the expansion in such cylinder. In 
the design of modern compound and triple compound engines we 
should start with the maximum already obtainable with a low-pressure 
cylinder ; that is, do as much work therein as has proved economical in 
low-pressure practice, then obtain as much work with the steam above 
that pressure as is practicable. The result will be that more work will 
be done in the low-pressure cylinder than in the high-pressure, as is, 
indeed, shown by the tests of the Leavitt engine, now under discussion. 
This does no harm. We have simply to provide for it in the design, 
even if two low-pressure cylinders are used, as in some forms of com- 
pound engine. 

I wish to thank Mr. Rockwood for the very earnest work he ha-> 
done in developing this question of compound vs. triple compound 
engines, though I do not think he is right in making such an extremely 
large ratio of capacity between the high and low-pressure cylinders. It 
is also a source of gratification that even better economical results have 
been obtained with a Leavitt compound engine, and as the latter result 
was secured with a less number of expansions, and with a larger pro- 
portion of the work done in the large cylinder, it indicates the correct- 
ness of the principles above stated. 

The general conclusion appears to be that we cannot as yet carry 
the steam pressure high enough to make the triple compound en. me of 

Trials of a Recent Compound Engine. 


value in a commercial sense. It is true that the best triple compound 
engines have given- a little better results than the best compound 
engines, but fairly large percentages of gain are for such economical 

ngines very small quantities, and are easily wiped out by very trifling 
d'( rangements, such as small leaks, want of care with jackets, etc., and 

re readily balanced by other items of cost, such, for instance, as a 
little higher wages of the engineer or the greater interest due to in- 
creased first cost. The coal is only one of the several items of cost in 
operating a steam-engine, and all must be considered in making a com- 

m ;ial balance. 

In making tb e remarks I wish to encourage rather than hinder 

any attempts to obtain the very 1 t results | ble. The chairman 
will realize that the anthr ite supply is limited ami that that kind ol 
coal will appreciate in pri. o the very extreme onomies will be 
valuable in the future, e n if not warranted by conm. ial considera- 
tions in the present. 

This discussion oi ompound vs. triple compound te will 

more valuable than seem it first sight. I hav< tin called tl 

attention of the Ameri< ua Society of Naval Engin to the sub; 
with a view of saving the weight and to some extent the 5] 
pied by the intermediate cylinder on board ship. Tl tl elemenl 
of space and displa* nent are of the great. importani ind, more 
ov he full power runs are compara vely so short thai -m 
can be sacrificed undei Juch circum nces, if economical results 
obtainable at ordinary cruising S] ds. It is true that tfa e-< rant, 

system of the tnple « , .pound engine i desirable feature i produi 
L smoot of working, but it m I not be ■ rificed if a return be 

made to compound engines, two of tl cylinderscan b 
cylinders, indeed, a common practice when lowei am | sun 

xs re employ The system of doing more work in tl rger , Iv.n- 

ders, previously recommended, aids m e solution of . 
though doubtless the ill be some difficulty m dtttebuti the t 

equally to the three cylinders. Th tern ada, itself very well b 
the condition, of varying loads obtainin m b rd sh.j md I have no 
doubt that in due tune valuable developments will \ mad, m tin 

direction. f . . . , 

Mr William Kent.-W- President, I think that ... ye 

engineers will read with great pleasure this , ,er Mr. 1 >ean s and 
the discussion by Mr. RocV -od, supplemented b ; Mr. 1 aery , 5 - 
cussion As the matter stands now we can hare learned ,t 

,e Leavitt engine, ae ling h Mr. Dean, is about five per c 
superior in economy to the Natick engme, and accordmg to Mr. 

58 Trials of a Recent Compound Engine 

Rockwood, if we make the proper allowances, the Natick engine is 
eight per cent, better than the Louisville engine. Add these figures 
together and divide them by two and we have the two engines very 
nearly alike. Mr. Rockwood mentioned in his remarks the Pawtucket 
engine, and it was also in my own mind at the time, as to what is the 
cause of the difference in economy between the Louisville engine and 
the Pawtucket engine. The Pawtucket engine had 16 expansions as 
against 20 in the Louisville and 33 in the Rockwood engine. The 
Pawtucket engine had 120 pounds of steam pressure as against 151 in 
the Louisville and 175 in the Natick engine. The Louisville engine 
had high vacuum as compared with the Pawtucket. The Pawtucket 
engine had only 240 for piston speed as compared with 371 for the 
Louisville and 611 for the Natick. It is probable that if the Pawtucket 
engine had been given 150 or 17; pounds pressure of steam and if the 
expansions had been 20 or 33 instead of 16, it would have shown a 
better result. So that the Pawtucket engine might stand pretty near 
the top if you would only give it the advantage these other engines had 
in steam pressure and expansion and piston speed. We cannot make 
a satisfactory comparison between the Natick and the Louisville 
engines, because the conditions are so different. The Louisville 
engine had 20 expansions. Was that the best practice for that particu- 
lar engine? The Natick engine had 33. Was that the best expansion 
for that particular engine? The steam pressure of the Natick engine 
was 175 against 150 in the Louisville engine. Were these pressures the 
best for those particular engines? Of course, t he vacuum in the Louis- 
ville engine was best, at all events^ and the Natick engine would, no 
doubt, have been glad to get such a vacuum. But I say we cannot 
satisfactorily determine which is the best form of these two engines 
until they are both tested with the same pressure and vacuum, 
and until each engine is tested with varying expansions, until they find 
the expansion best suited for that particular engine. 

In regard to Mr. Emery's remarks, he makes a point about the 
compound vs. the triple-expansion engine for marine practice It is 
strange that, about 1882, the very engine he speaks of, the compound 
engine with three cylinders, was the favorite engine, and it has paid 
since that to take these out of the ships and substitute the triple-expan- 
ion at great cost, a great economy resulting from the change, although 
I admit that putting in boilers of higher steam pressure might have 
been largely the cause of the economy. We cannot determine that, 
however, because we have not had a trial of that particular form of 
compound engines with high pressure and with moderate pressure 
steam. We do not know to-day what that engine might have been 

Trials of a Recent Compound I gini 


capal of doing with steam of one hundred and seventy-five pounds 
pressure, because it never was tried. 

Mr, R. S. Hale.— I should like to ask Mr. Dean what w 
slip of the engine, as determined by weir m< urement. I um r 

Mr. Braekett spoke to me of thing like seven per cent, tnd il 

was as mm h is that, would it not change considerably th friction of 
the engin. iul the duties, as figured, of tl plunger disp nm 

Mr j v/ ._i would like to a^k lor in rmation with fer- 

ll(t . t0 ti boiler practici it this mill engii The . ttr 

neenth m\ I of five how nd one hundred and forty-four 1 
witboul anyexpl m with ret- nee to the boiler pi jms 

t,, me to leas- ow thing la< king. 

Mf Dean.— The test if tl ;in< ■ are reported « 

simpl d-wat. it, and the water in rs v. the 

I,, at the end ol the test as the 1 innii wd th. mi pr< 
sur e was tl lw rhat, with a little exp m always 

brought al 

That par' >1 the I I of tl l >u« engii re to »j Mr. 

Hale w »t touched upon b me in tl »rther ml I 1 

more int. I in th. aa | rfonnam I ". an) r | 

though! prol othei , ipl ou Id also. Tl lip 

engine is remarkably larg It n • lar th it I k u 
orthT , to try to find out the n on md it 
percenl We di mined the data for i da h. 

t different cuti ; that is t. y t I iusi 

mined Lb. lata ich daj separately, and it . alw 
We were some* auspicious o ur * But all ol ipi ; 

so , wt tt able ti , proved to be groundl W 

times stop] I the engine, and shut vab in the main, am 

flovv f the v in the chamber at ti 

when the val shut, that it all cam. 

tt such tii the How wate about one ] I * 

of the plunger pfe nt. Th ,nly » tl 

this unusual di] i th -that the pu, i ^ 

seat i n a metallic * and the Ohio River water 

sand, nit. d those v ah , in 

or less, and val hat were taken on med 

were on on ude, and not all th. md. I 

hardly imagine a pump-valve ^here ,n: 

general thing, to , down, and thai U - M 

, , nc ,d tlittl picker than does l .ther and . 
pump-val i must be 1 el) fitted so that the) fr 

6o Trials of a Recent Compound Engine. 

under all conditions. It looked as if the valve in general struck an 
the edge and gouged the seat out, so that we thought that probably a 
good deal of the slip was to be accounted for in that way: and in 
listening to the pumps, putting one's ear right against the pump 
chambers, there was a sound which did indicate that there was water 
going through somewhere — it was rather difficult to tell where — at the 
time when the water was being forced op into the main. Hut I do not 
know that I can throw any additional light on that subject. The 
whole matter wa> an astonishment to all of us, and we u-ed a good 
deal of time to try to find out what the trouble was. That, however 
would not, as Mr. Hale suggests, affect the friction of the engine. 

Now that I am on my feet I will speak of some other interesting 
things which were done with that engine, and which are not mentioned 
in the paper. The result was so unusual that I thought I wou'd go to 
rather unusual pains to corroborate it, and in the report — I have for- 
gotten whether I stated in this paper or not — but in the report of the 
test it is mentioned that the condensation in the jacket was determined 
by passing it through a Worthington meter, which meter worked with 
remarkable steadiness. It always showed about twenty-five cubic units 
per hour; whether you took the data on the first day, or third day, or 
last day, it was just the same. Immediately after the trial I calibrated 
that meter for some three hours. I was hardly content, however, with 
that calibration, and after I got home I wrote to the chief engineer of 
the waterworks to ask him to determine that condensation for me l>\ 
actually weighing the jacket condensation, and also to run another test 
of twentv-four hours' duration; and I will say here that Mr. Herman} 
had a very competent chief assistant, who helped me in this test and in 
whom I had the utmost confidence. He fully appreciated the nece--i- 
ties of the case. Persons who have read the account of the test in the 
report will remember that the amount of water by the feed-pump was 
determined by computing it from the rise of temperature of the water 
before it w r as heated by this pump exhaust, and after. But in this sup- 
plementary test which I asked Mr. Hermany to make, the exhaust vva 
turned out of doors and the feed-pump was run by the donkey boiler 
and the jacket water was actually weighed throughout the twenty-four 
hours, and the separator condensation also. This jacket condensation 
differed from that which I had determined by .06 of one per cent 
The head of water on the pump was almost identical, the revolutions 
were just the same, and the indicated horse-power figured out precise! \ 
the same as on the official test. On each of the six davs of the test 
the amount of feed-water used by the engine was 187,000 pound-, 
almost without exception. It differed only a small number of pound> 

Trials of a Recent Compound Engine. 


The greatest difference that we found from my results was the separa- 
tor condensation for the twenty-four hours. I made it 3,900 pounds 
in twenty-four hours, and he made it 2,800. There was a difference, 
you see, somewhere about one-half of one per cent, of the total feed. 
We are dealing with such large quantities that it is of no importance. 
He also ascertained for me the two jacket condensations separately, and 
the re-heater condensation separately, but simultaneously. All of the 
data which are given in my report have been so thoroughly corrobo- 
rated and reproduced day after day on that test that they are singularly 

to be relied upon. 

^Replying to Mr. Ball.— As to the effect of compression on 
economy, the experiments to which he refers as having been carried 
out by Professor Jacobus were made on a relatively low-grade engine. 
By that I mean a single-valve engine with large clearances. Results 
from such an engine, I believe, are little or no guide in determining 
practice with high-grade engines. By high-grade engines I mean four- 
valve engines with small clearances. With low-grade engines some 
thermo-dynamic phenomenon with high compression may creep in 
which overpowers others. In the high-grade engine there is less room 
for erratic phenomena, and we can work more closely to our theories 
and obtain corresponding results. The Leavitt engine 1. worked out 
in detail close to the theories, and the results are given in my paper. 
Mr. Ball's arguments do not appeal to me, either with reference to 

compression or to drop. 

With reference to drop, I will simply say that the modern engine 
is made to use steam expansively. It may be done in one cylinder, 
but it has been found that it is much more economical to divide it up 
into steps, each cylinder performing a step. Why should not one step 
begin where the preceding one leaves off? I confess that I never have 

been able to see. 

As I understand it, Mr. Ball claims advantage in drop, because it 
superheats the steam. If we assume steam of 45 pounds absolute to 
drop to 25 absolute, and thus to drop 20 pounds, the superheat will be 


- 1I 55- I _ 2I g ? o This superheat would not, however, exist, 

for the 'released heat would find itself in wet steam, and therefore the 

supposed benefit is all but nil 

The amount of heat added to a pound of steam of the lower pres- 
sure would be ««5.*-«5M=»"-S B. T. U., or A of 1 per cent., 
and this, in turn, would dry out 10.5-^922 = 0.013- or i^ per 

* Author's closure, under the Rule 


62 Trials or a Recent Compound Engine. 

cent, of moisture in the steam, the benefit of which is unknown, but 
small. In order to secure this small benefit Mr. Ball would lose ex- 
pansive energy of the steam the value of which is exactly known, and 
is represented by 1 + hyp. log .* O- = 1 + hyp. log. 1.8=1.5878 per 
pound of steam. I prefer to get this work out of the steam, especially 
when its quality is restored by a re-heater. 

Replying to Mr. Rockwood ; I think it is not unreasonable to sup- 
pose that the whole engine at Natick was built in accordance with the 
Rockwood system, and therefore to be criticised as such. I am some- 
what in sympathy with Mr. Ball in not understanding what the Rock- 
wood stem is. If it is a ratio of 7 to 1 it is seldom made by him. 
and has in a general way been put into English steamships several 
years since. As for tests of triple engines with the intermediate cylin- 
der cut out, the columns of Engi?ieering contain the results of tests 

that are disastrous to the cut-out. 

Mr. Rockwood states that the steam -pipe leaked at the joints at 
Natick. This is not true, except at a slip expansion joint, and the 
amount was so slight as to make no perceptible difference in the result. 

The shortness of the tests of the Natick engine are unfavorable to 
it compared with the Louisville. 

I agree with Mr. Rockwood that, so far as I know, the advantage 
of high piston speed is mythical, and cuts but little figure in the com- 
parison made by me. 

The probable greater economy that would be due to a better 
vacuum with the Natick engine was estimated by me by ascertaining 
how much more area would have been given to the L. P. indicator 
diagram thereby, and the resultant increase in work done by the steam. 

Mr. Rockwood entirely misinterprets the sag in the exhaust line 
of the H. P. indicator diagram of the Leavitt engine, except that it 
increases the range of temperature in that cylinder. He, however, 
makes no allowance for the fact that this increase in range is corrected 
by the rise in pressure in the receiver and H. P. cylinder after th- 
closure of the L. P. inlet valve. This correction is intentional, both 
to avoid drop in pressure and net drop in temperature. 

To make myself understood with reference to drop I will define 
it. In general, drop is a fall in pressure between the end of expansion 
in one cylinder and the beginning of expansion in the next, and, 
specifically, in a tandem or Leavitt engine, it is the fall in pressure be- 
tween the terminal pressure in one cylinder and the initial pressure in 
he next. The only unavoidable drop in such engines is due to the 
work of moving steam from one cylinder to the other. In the Leavitt 
engine, expansion in the second cylinder begins with the beginning of 

Trials of a Recent Compound Engine. 



the movement of the piston from the end of the stroke, and no fall in 
pressure can occur until piston movement begins ; while drop is unre- 
sisted expansion, or fall in pressure without piston movement and with- 
out doing work, and is therefore a dead loss, except in so far as the heat 
released produces some superheating. It must not be overlooked that 
true expansion takes place with the whole stroke of the Leavitt L. P. 
piston, only that the law changes after L. P. cut-off. 

Mr. Rockwood is wholly wrong in supposing that a larger receiver 
would produce drop in the release end of the high-pressure diagram. 
This has nothing whatever to do with it, as a drop, or its absence, 
will be determined by the low-pressure point of cut-off in either a 
tandem, Leavitt, or cross-compound engine. If valves are properly 
set in either of these types of engine, and the cut-offs of all but the 
first cylinder are not affected by the governor, and a permanent regime 
has been established, neither will ever produce drop or loops in the 
indicator diagrams, except the always unavoidable drop above men- 
tioned, and which increases with speed. This is furthermore entirely 
independent of the receiver volume, or point of cut-off in the high- 
pressure cylinder. The large receiver will diminish the temperature- 
range in the high-pressure diagram, and is so far beneficial unless the 
correction above referred to is wholly effective. It will, however, not 
affect the range of temperature in the low-pressure cylinder, as Mr. 
Rockwood claims, because the initial and back pressures in this cylin- 
der are not affected by the receiver. 

My understanding of the effect of equal ranges of temperature in 
cylinders is not, as Mr. Rockwood says, that "an equal amount of 
cylinder condensation will occur," but that a minimum total conden- 
sation will occur. Although I cannot now give an absolute proof of 
this, I am satisfied to hold this view for the present. The theory that 
Mr. Rockwook tentatiously advances, viz., that equal range takes no 
account of the amount of cylinder surface, and that the large cylinder 
would necessarily condense much more than the small, is incontinent 
with facts, for we know that in every engine the condensation is 
greatest in the small cylinder. 

Finally, after all has been said and written, the fact remains that 
an engine with a cylinder ratio of 4 to 1 has surpassed in economy an 
engine with 7 to 1, carrying a higher steam pressure.